import os, sys from datetime import datetime from dateutil import relativedelta sys.path.append('/var/www/digisnaxx.ado/scrapers') import dtss dtss.getReady() from time import sleep from pprint import pprint as ppr import pytz from events.models import Organization, Scraper, Calendar, Event import events.digitools as digitools tz_str = "-0500 UTC" DATETIME_FORMAT = '%A, %B %d %Y %I:%M %p %z %Z' venue, created = Organization.objects.get_or_create( name="JP Centre Yoga", city="Boston", website="https://www.jpcentreyoga.com/yoga-class-schedule", is_venue = True ) scraper,item_count_start,virtcal = digitools.getScraper(venue, venue.website, 'bos') scraper.items = 0 scraper.save() def get_events(ps, event_type): contents = ps.xpath('.//*/div[@class="bw-widget__day"]') for cn in contents: date = cn.xpath('.//div[1]/text()')[0] events = cn.xpath('.//div[@class="bw-session"]') for c in events: try: event = {} time = c.xpath('.//*/time[@class="hc_starttime"]/text()')[0] event['scraper'] = scraper event['calendars'] = [scraper.calendar] event['title'] = c.xpath('.//*/div[@class="bw-session__name"]/text()')[1].replace("\n", "").strip() year = datetime.now().year event['date'] = [date, str(year), time, tz_str] event['date'] = " ".join(event['date']) event['dateStamp'] = datetime.strptime(event['date'], DATETIME_FORMAT) event['link'] = venue.website event, created = digitools.createBasicEvent(event, "Ex", venue) if "o n l i n e" in event.show_title: event = digitools.add_calendar(event, "000") scraper.items+=1 except Exception as e: print("\nError: ", e) pass if len(sys.argv) >= 2: arg1 = sys.argv[1] br = digitools.getBrowser(arg1) else: print("No run_env") br.close() quit() ps = digitools.getSource(br, venue.website) get_events(ps, "Co") sleep(3) digitools.updateScraper(scraper, item_count_start) br.close()